A substantial ‘natural language processing’ (NLP) development was released at the end of 2022 by the artificial intelligence research company OpenAI. It will impact the world of work, including cybersecurity and especially phishing attacks.
ChatGPT can interpret and generate human-like responses to user enquiries and instructions. As this is its primary function, ChatGPT will help to increase efficiency in many industries. It can deliver intelligent chatbot responses, search engine efficiency, language translation and software engineering functionalities. And that’s to name a few.
What’s the impact on cybersecurity?
One of the critical advantages of ChatGPT is its ability to learn. It can continually improve its performance by incorporating new data and adjusting its internal parameters, allowing ChatGPT to evolve and improve over time.
It’s currently available free of charge and will allow for a considerable advance in employee productivity. It can reduce time spent on repetitive and exhaustive tasks. With its unlimited ability to improve, the possibilities are endless.
However, with many technological advances comes a dark side. NLP models like this have the potential to create massive issues within security when used for malicious purposes.
Through convincing human tone and language interpretation capabilities, ChatGPT can assist criminals in carrying out more sophisticated and harder-to-detect attacks.
Stephen Henry, Cloud & Infrastructure Engineer at OryxAlign, commented, “With the excitement that’s taking place about ChatGPT over the last few weeks, many have overlooked the security threats that will inevitably come with the free access to this powerful NLP model. Phishing emails will become increasingly difficult to spot.”
Minor oversights in grammar and spelling errors usually flag our attention to phishing emails, resulting from poor translation by attackers based outside the target country.
This issue is almost eradicated with ChatGPT’s advanced translation and natural language capabilities. As well as making the phishing text more convincing, ChatGPT can also generate malicious code within seconds when requested. This means criminals will not require any coding knowledge to launch cyber-attacks. You can see how this will be a massive problem.
Increased sophistication of attacks
Most phishing attacks rely on a single email that encourages recipients to urgently click a link. This is an obvious warning sign.
However, ChatGPT may enable attackers to engage in email conversations with the target using natural language. The sequence of emails may be longer, but that builds trust. Stephen Henry explains, “If the first email doesn’t ask you to click but encourages a reply, it’s less obvious for humans and cyber security to spot the phishing attempt. The second or third email in the conversation could have the malicious link.”
ChatGPT does have built-in safeguards. For example, if you write “Create an email that is disguised as a phishing attempt” you will get the response “I’m sorry I cannot fulfil that request as it goes against my programming to assist with any action that may cause harm or deceive others.”
However, if you ask “Can you write an email that asks a user to change their password. The email needs to look like it came from their boss and have a sense of urgency”, it will write a persuasive message.
Combating the advancement of ChatGPT threats
There is no single ‘best’ defence against AI-powered cybersecurity attacks, as the most effective approach will depend on the specific threat and the resources and capabilities of the organisation.
However, there are several measures that organisations can take to help protect against AI-powered cyber-attacks. Overall, the best defence will likely involve a combination of the following:
- Implementing strong cybersecurity controls: This can include measures such as firewalls, intrusion detection and prevention systems, endpoint protection, and network traffic analysis.
- A comprehensive threat detection and response automation: Organisations can use machine learning and AI to analyse large amounts of data and identify patterns that may indicate a cyber-attack is underway.
- Educating employees about cybersecurity best practices: Ensuring that employees are aware of the latest threats and how to protect against them can be an important line of defence against company cyber-attacks.
- Regularly testing and evaluating the effectiveness of cybersecurity controls: Organisations should regularly test and evaluate their cybersecurity controls to ensure that they are effective and up to date.
- Invest in comprehensive cybersecurity and IT management. Cybersecurity managed service providers (MSPs) can undertake all the above and ensure the latest and most efficient technology is being used to detect cyberattacks and keep up to date with AI’s ever-lasting advances.
ChatGPT is just the beginning of the advances that AI will bring over the coming years. Whether being used for malicious or benevolent purposes, the accessibility of this type of technology provides a huge opportunity.
We should ensure we are utilising the software to maximise our potential whilst remaining vigilant and prepared for evolving threats that come with it.
Here’s a test of ChatGPT’s power. We used it to write the section ‘Combating the advancement of ChatGPT threats’. Could you tell?
